Estimation of forest tree types structure by using classification of airborne lidar data a case study: Istanbul, Belgrad Forest

Airborne lidar is a progressively used technology for land survey applications as, 3D city planning and modeling, forest survey and geomorphologic structure determination etc. Airborne Lidar systems basically consist of laser scanners, global positioning system (GPS) and Inertial Measurement Unite (IMU). Since 2003, forestry studies are in progress in industrial countries, however in Turkey first large scale airborne lidar project is Nabucco (2009-2011). In this study, classification of high density forest (55points/m2) and the possibility to predict tree height, above ground tree species biomass and canopy cover are investigated with airborne laser scanning (ALS) data. With calibrated lidar data, some classification parameters were used on TerraScan software. First of all, to avoid confusion, water areas were removed. Through multiple echo (discrete echo) information and "ground classification" which uses Axelsson algorithm, lidar data has been separated into six classes as well as in "American Society for Photogrammetry and Remote Sensing" ASPRS classification list (low points, ground, low-medium-high vegetation, water). The most difficult step was to separate ground and low points, for this step automatic and also manual corrections were done. After all, some forest tree structure was detected with cross sections and was searched tree types of forest. 6 tree types have been calculated the numbers of trees were calculated approximately. Moreover, scan angle comparisons were done with different lidar point data and the statistical results about the tree classes were calculated. All analyses were done with Microstation, TerraScan, Matlab and Surfer software. As a result, it is concluded that airborne lidar is not only dependent on different paramaters in lidar system hardware but also software and algorithms.